Multi-category driving maneuver recognition based on smartphone sensors and XGBoost

Fujian Wang; Hui Tian; Shengqiang Jia; Zhenyu Mei

doi:10.1117/12.3030478

16 May 2024 Multi-category driving maneuver recognition based on smartphone sensors and XGBoost

Fujian Wang, Hui Tian, Shengqiang Jia, Zhenyu Mei

Proceedings Volume 13160, Fourth International Conference on Smart City Engineering and Public Transportation (SCEPT 2024); 131600W (2024) https://doi.org/10.1117/12.3030478
Event: 4th International Conference on Smart City Engineering and Public Transportation (SCEPT 2024), 2024, Beijin, China

Abstract

Driving maneuver recognition is one of the important problems in the field of intelligent transportation and is the basis of driving behavior research. The data collected from the accelerometer, gyroscope and GPS of the smartphone during natural driving were processed to obtain 14 types of driving maneuvers, including lane changing, obstacle avoidance, overtaking, 45° turning, 90° turning, 180° turning, noise, and idle (where noise refers to the noise data received while using the smartphone), and establish a driving maneuver sample dataset, where the distribution of sample categories is extremely imbalanced, the idle class with the most samples accounts for 97.006%, while the 180° right turn class with the least samples only accounts for 0.002%. So this paper proposes a driving maneuver recognition model based on XGBoost (eXtreme Gradient Boosting) algorithm, which utilizes the XGBoost algorithm to extract and analyze features from the sample dataset without artificially balancing the number of samples in each category, and the test results show that the recognition accuracy is 0.997, and the macro averages of precision, recall, and F1 score are 0.982, 0.956, and 0.968, respectively, which are significantly better than that of Random Forest, Adaboost and LightGBM, and overall better than that of CNN (Convolutional Neural Network) and LSTM (Long Short Term Memory). Therefore, XGBoost has excellent driving maneuver recognition capability, which can solve the problem of low recognition rate due to the extreme imbalance in the number of samples of each category in the multi-category driving maneuver recognition task.

(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Fujian Wang, Hui Tian, Shengqiang Jia, and Zhenyu Mei "Multi-category driving maneuver recognition based on smartphone sensors and XGBoost", Proc. SPIE 13160, Fourth International Conference on Smart City Engineering and Public Transportation (SCEPT 2024), 131600W (16 May 2024); https://doi.org/10.1117/12.3030478

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available

Members: $17.00

Non-members: $21.00 ADD TO CART

PROCEEDINGS
8 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Machine learning

Sensors

Data modeling

Deep learning

Performance modeling

Accelerometers

Gyroscopes

Keywords/Phrases

Search In:

Publication Years